REPORTS
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
2378
samples (from 2- and 10-m depth) for DIC, Alk, and
␦13C at Station S, located 26 km southeast of the
island of Bermuda (32°10’N, 64°30’W ) (27). In 1989,
the Joint Global Ocean Flux Study BATS program was
established at a station 56 km further southeast
(31°50’N, 64°10’W ) (38). The CDRG program extended their measurements to this new site, and the
BBSR started to measure DIC there throughout the
water column (24). Because differences between Station S and BATS have been found to be small (38),
and because the DIC data from the two laboratories
show no systematic differences (change in DIC ⫽
⫺1.0 ⫾ 2.2 ␮mol kg⫺1, N ⫽ 78 measurements), the
data from the two sites and two laboratories are
combined here into a single time series.
The reduced isotopic ratio, ␦13C, is defined as ␦13C ⫽
(13rsample – 13rstd)/13rstd, where 13rsample is the
13C/12C ratio of the sample, and where 13r
is the
std
13C/12C ratio of the Pee Dee belemnite standard.
We computed the oceanic pCO2 from the observed
temperature, salinity, Alk, DIC, and nutrients using a
thermodynamic model of the carbonate system (39)
and the dissociation constants of Mehrbach et al. (40)
as refitted by Dickson and Millero (41). This choice is
based on laboratory studies (42) and the resulting
good agreement with direct observations of pCO2
near Bermuda (43).
R. B. Bacastow, C. D. Keeling, T. J. Lueker, M. Wahlen,
W. G. Mook, Global Biogeochem. Cycles 10, 335
(1996).
N. Gruber et al., Global Biogeochem. Cycles 13, 307
(1999).
N. R. Bates, Deep-Sea Res. II 48, 1507 (2001).
This trend is smaller than that reported by Bates (24)
for 1988 to 1998, mainly because of smaller growth
rates in the 1980s.
R. E. Sonnerup et al., Global Biogeochem. Cycles 13,
857 (1999).
C. D. Keeling, in The Global Carbon Cycle, M.
Heimann, Ed. (Springer-Verlag, New York, 1993), pp.
413– 430.
T. M. Joyce, P. Robbins, J. Clim. 9, 3121 (1996).
N. Gruber, C. D. Keeling, T. F. Stocker, Deep-Sea Res.
I 45, 673 (1998).
Net community production refers to the net transfer
between inorganic and organic carbon pools due to
photosynthesis and to oxidation of organic matter,
and therefore is equal to net primary production
minus community respiration.
Materials and methods are available as supporting
material on Science Online.
Because the time-series sites are located in the broad
recirculation region of the subtropical North Atlantic,
the geostrophic mean currents are from the northeast. Current meter data indicate a mean current of
about 0.05 m s⫺1 (44), although substantially higher
velocities are associated with the passage of eddies.
Based on the observed mean horizontal gradient near
Bermuda of about 1.1 ⫻ 10⫺5 ␮mol kg⫺1 m⫺1 (29),
our model-derived transport estimates give horizontal velocities of between – 0.01 and – 0.20 m s⫺1,
with a long-term mean of – 0.07 m s⫺1 (negative
velocity indicates southward flow).
The uncertainties of the fluxes have been established
with a Monte Carlo technique (45). The diagnostic
model was run 1000 times, with randomly selected
sets of parameter values within a predetermined
range for each parameter based on the parameter’s
perceived uncertainty (29). The flux uncertainties
listed in the text denote the 1-␴ uncertainty computed from the results of these 1000 realizations.
Because our diagnostic model analyses cannot
uniquely determine the underlying mechanisms, other factors, such as variations in mesoscale eddy dynamics, or dust deposition and its possible impact on
N2 fixation, could be responsible for our positive
correlation of net community production with winter
mixed-layer depth anomalies.
R. L. Molinari, D. A. Mayer, J. F. Festa, H. F. Bezdek, J.
Geophys. Res. 102, 3267 (1997).
N. R. Bates, A. C. Pequignet, R. J. Johnson, N. Gruber,
Nature 420, 489 (2002).
The atmospheric CO2 inversion results have been
provided by P. Bousquet and P. Peylin and are based
38.
39.
40.
41.
42.
43.
44.
45.
on the anomalous CO2 fluxes for the entire North
Atlantic region as estimated in their global study (4).
A. F. Michaels, A. H. Knap, Deep-Sea Res. II 43, 157
(1996).
U.S. Department of Energy, “Handbook of methods
for the analysis of the various parameters of the
carbon dioxide system in sea water,” version 2, Tech.
Rep. ORNL/CDIAC-74 (U.S. Department of Energy,
Carbon Dioxide Information Analysis Center, Oak
Ridge National Laboratory, Oak Ridge, TN, 1994).
C. Mehrbach, C. H. Culberson, J. E. Hawley, R. M.
Pytkowicz, Limnol. Oceanogr. 18, 897 (1973).
A. G. Dickson, F. J. Millero, Deep-Sea Res. 34, 1733
(1987).
T. J. Lueker, A. G. Dickson, C. D. Keeling, Mar. Chem.
70, 105 (2000).
N. R. Bates, T. Takahashi, D. W. Chipman, A. H. Knap,
J. Geophys. Res. 103, 15567 (1998).
D. A. Siegel, W. G. Deuser, Deep-Sea Res. I 44, 1519
(1997).
R. Y. Rubinstein, Simulation and the Monte Carlo
Method (Wiley, New York, 1981).
46. T. J. Conway et al., J. Geophys. Res. 99, 22831 (1994).
47. We are grateful to the numerous people who have
been responsible for the collection, preparation, and
analysis of the data presented here. We thank in
particular P. Guenther, T. Lueker, G. Emanuele, E.
Bollenbacher, and A. Dickson at the Scripps Institution of Oceanography. We also thank A. Knap, A.
Michaels, D. Steinberg, C. Carlson, and D. Hansell for
their help and support. We are grateful to P. Bousquet and P. Peylin for providing us with their atmospheric CO2 inversion results for the North Atlantic.
Supported by NSF grants OCE-0097337 (N.G.) and
OCE-0083918 (C.D.K.) and by grants from the NSF
and NOAA (N.R.B.).
Supporting Online Material
www.sciencemag.org/cgi/content/full/298/5602/2374/
DC1
Materials and Methods
Table S1
References and Notes
7 August 2002; accepted 13 November 2002
Control of Facial Muscle
Development by MyoR and
Capsulin
Jian-rong Lu,1 Rhonda Bassel-Duby,1 April Hawkins,1
Priscilla Chang,1 Renee Valdez,1 Hai Wu,1 Lin Gan,4
John M. Shelton,2 James A. Richardson,1,3 Eric N. Olson1*
Members of the MyoD family of basic helix-loop-helix (bHLH) transcription
factors control the formation of all skeletal muscles in vertebrates, but little is
known of the molecules or mechanisms that confer unique identities to different types of skeletal muscles. MyoR and capsulin are related bHLH transcription factors expressed in specific facial muscle precursors. We show that
specific facial muscles are missing in mice lacking both MyoR and capsulin,
reflecting the absence of MyoD family gene expression and ablation of the
corresponding myogenic lineages. These findings identify MyoR and capsulin as
unique transcription factors for the development of specific head muscles.
The myogenic bHLH transcription factors—
MyoD, Myf5, myogenin, and MRF4 —control vertebrate skeletal muscle development
(1). MyoD and Myf5 act redundantly as myoblast specification genes (2), whereas myogenin is required for myoblast differentiation
(3, 4). MyoD and MRF4 also play redundant
roles in muscle differentiation (5). Although
these myogenic genes control a developmental program shared by all skeletal muscles,
there is evidence that muscles in the head and
trunk differ with respect to the early steps in
myogenic lineage specification (6). For example, in Myf5⫺/⫺Pax-3⫺/⫺ double-mutant
mice, the trunk musculature is eliminated, but
head muscles are unaffected (7). The devel1
Department of Molecular Biology, 2Department of
Internal Medicine, and 3Department of Pathology, The
University of Texas Southwestern Medical Center at
Dallas, 6000 Harry Hines Boulevard, Dallas, TX
75390 –9148, USA. 4Center for Aging and Developmental Biology, University of Rochester, Rochester,
NY 14642, USA
*To whom correspondence should be addressed. Email: eolson@hamon.swmed.edu
opmental control genes responsible for head
muscle formation are unknown.
All skeletal muscle posterior to the head is
derived from paraxial mesoderm that becomes segmented into somites (1, 8). In contrast, head muscles are derived from multiple
cell lineages, including prechordal mesoderm
anterior to the first somite and paraxial mesodermal precursors that migrate into the
branchial arches (9–11). MyoR/musculin and
capsulin/Pod-1/epicardin are related bHLH
proteins transiently expressed in migratory
paraxial mesodermal cells in the branchial
arches that appear to represent precursors of
the muscles of mastication (12–18).
A null mutation in the mouse capsulin
gene results in neonatal lethality due to
pulmonary hypoplasia, but no overt skeletal
muscle abnormalities (18, 19). We combined mutations in MyoR and capsulin to
see if there were shared activities or effects
not revealed with either single-gene deletion. The MyoR gene was targeted in embryonic stem (ES) cells by homologous
recombination (fig. S1A). ES cell clones
20 DECEMBER 2002 VOL 298 SCIENCE www.sciencemag.org
REPORTS
harboring the mutant allele were used to
generate chimeric mice, which transmitted
the mutation through the germ line (fig.
S1B). Breeding of mice heterozygous for the
mutant MyoR allele yielded homozygous mu-
tants at predicted Mendelian ratios with no
obvious abnormalities.
Mice homozygous for the MyoR and capsulin null mutations were obtained by breeding heterozygous mutant mice. Double mu-
Fig. 1. Deficiency of head skeletal muscle and diaphragmatic hernia in MyoR⫺/⫺capsulin⫺/⫺
neonates. Coronal sections from MyoR⫺/⫺capsulin⫹/⫺ (A and B) and MyoR⫺/⫺capsulin⫺/⫺ (C and
D); (B and D) show a higher magnification of (A and C), respectively. Note the cleft palate ( p) in
(C). gl, glands; m, mandible; ma, masseter muscle; p, palate; pt, pterygoid muscles; t, tongue; te,
temporalis muscle. Asterisks in (C) denote missing muscles. In (B), the glands abut the masseter,
whereas in (D), the masseter is missing and the glands abut the mandible. (E) Diagrams of muscle
groups missing from the double mutant. (F) Sagittal section of a MyoR⫺/⫺capsulin⫺/⫺ with
diaphragmatic hernia. Arrowheads mark the boundaries of the diaphragmatic defect. Arrow marks
the diaphragm. d, diaphragm; g, gut; l, liver; p, pancreas. Scale bars, 200 ␮m.
tants were born alive, but, like capsulin⫺/⫺
mice (18, 19), they died within minutes after
birth. Histological examination revealed a
complete absence of the major muscles of
mastication, including the masseter, medial
and lateral pterygoids, and temporalis muscles in the majority of double-mutant embryos (Fig. 1, A to E). In their place was connective tissue. In a subset of double mutants,
atrophic pterygoid myofibers persisted unilaterally. The missing muscles in MyoR⫺/⫺
capsulin⫺/⫺ double-mutant mice are derived
from the first branchial arch and represent a
distinct group of muscles that function in
mastication (8). Other first arch– derived
muscles, such as the anterior digastric and
mylohyoid, which do not function in mastication, were present in the double mutants
[Fig. 1, A to D and (20)]. Trunk muscles of
double mutants were also indistinguishable
from those of wild-type animals (20). Head
muscle defects were not observed in
MyoR⫹/⫺capsulin⫺/⫺ or in MyoR⫺/⫺capsulin⫹/⫺ embryos. These findings demonstrated
that MyoR and capsulin redundantly controlled the formation of a specific subset of
first arch– derived facial skeletal muscles and
that a single copy of either gene is sufficient
to support normal muscle development.
MyoR⫺/⫺capsulin⫺/⫺ double mutants
also displayed cleft palate (Fig. 1C). Other
branchial arch– derived skeletal elements
were normal in double mutants [Fig. 1 and
(20)]. In addition, the visceral organs of
MyoR⫺/⫺capsulin⫺/⫺ double-mutant neo-
Fig. 2. Expression of capsulin-lacZ allele. Expression of lacZ in
capsulin⫺/⫺ mice at E7.5 (A) and E8.0 (B and C), as indicated
by arrowheads. (C) A coronal section through the embryo in
(B). (D to F) LacZ staining is present within the first and
second branchial arches in MyoR⫺/⫺capsulin⫹/⫺. (G to I)
Staining gradually disappears from the first arch of
MyoR⫺/⫺capsulin⫺/⫺. (J and K) Sections of the first arch of
embryos in (F) and (I), respectively. 1, branchial arch 1; 2,
branchial arch 2. Arrowheads in ( J) and (K) indicate the muscle
developing around the fifth cranial nerve. LacZ-positive cells are
missing from this region of the double mutant. Scale bars: (A,
B, and C), 200 ␮m; (D to I), 500 ␮m; ( J and K), 100 ␮m.
www.sciencemag.org SCIENCE VOL 298 20 DECEMBER 2002
2379
REPORTS
Fig. 3. Expression of myogenic bHLH genes in wild-type and MyoR⫺/⫺
capsulin⫺/⫺ embryos. (A) MyoR, capsulin, and Myf5 expression were
detected in the branchial arches of wild-type embryos by in situ
hybridization. Capsulin was detected at E8.5 in the first arch (ba1).
Expression of Myf5 and MyoR overlapped that of capsulin at later
stages. (B) Myf5 expression was detected in the first branchial arches
of MyoR⫹/⫹capsulin⫹/⫺, designated wild type, but not in MyoR⫺/⫺
capsulin⫺/⫺ embryos. The middle panels show higher magnification of
the upper panels. ba1, first branchial arch; ba2, second branchial arch;
hc, hypoglossal cord. m, Myotome. (C) Myf5, myogenin, and MyoD
expression was detected in wild-type and MyoR⫺/⫺capsulin⫺/⫺ embryos at E15.5. The masseter muscle (ma) is completely absent in
MyoR⫺/⫺capsulin⫺/⫺ mutants (designated ma*). Arrow indicates cleft
palate. Scale bars: 200 ␮m for (A and B), 1 mm for (C).
Fig. 4. Apoptosis of
muscle cells in the first
branchial
arch
in
MyoR ⫺/⫺capsulin ⫺/⫺
embryos. First branchial
arch of MyoR⫺/⫺ capsulin⫹/⫺ and MyoR⫺/⫺
capsulin⫺/⫺ embryos at
E10.5 stained for lacZ
expression and viewed
by differential interference contrast microscopy. (A and B). The same
sections were stained
with propidium iodide,
and apoptotic cells were
identified by TUNEL labeling (C and D). At this
stage, lacZ-positive cells
are disappearing from
the first branchial arch
of the double mutant because of apoptosis, as revealed by TUNEL labeling (yellow) in the lacZ-positive cores.
LacZ staining interferes with detection of propidium iodide. Arrowheads point to lacZ-positive cores in the
first branchial arches. Scale bar, 100 ␮m,
nates were displaced into the chest through a
defect in the posterior region of the diaphragm (Fig. 1F). This abnormality suggests
that MyoR and capsulin are required to maintain integrity of the diaphragm muscle. The
diaphragm develops from fusion of the septum transversum, the pleuroperitoneal membrane, and the dorsal esophageal mesentery
and is populated by muscle precursor cells
that migrate from the cervical somites (21).
MyoR is expressed in diaphragmatic myoblasts (12, 13), and capsulin is expressed in
the septum transversum and the pleuroperitoneal membrane (14–18).
The head muscle deficiency in MyoR⫺/⫺
capsulin⫺/⫺ double mutants could arise from a
2380
defect in specification, migration, or proliferation of the affected myogenic lineages; a block
in myoblast differentiation; or an increase in
apoptosis. To determine the basis for this muscle deficit and to pinpoint the time of its onset,
we compared the expression of a lacZ reporter
integrated into the targeted capsulin allele in
staged embryos of the different genotypes.
From embryonic day 7.5 (E7.5) to E8.0, when
mesodermal precursors of first arch muscle
cells initially appear subjacent to the metencephalon (10), lacZ staining was localized to
this muscle precursor population in embryos of
the different genotypes (Fig. 2, A to C). At
E9.5, these lacZ-positive cells could be seen
migrating into the newly formed branchial
arches of wild-type and double-mutant embryos
(Fig. 2, D and G). By E10.5, a swathe of
lacZ-positive cells extended into the first and
second branchial arches. This myogenic precursor pool appeared to migrate properly, but there
were noticeably fewer of these lacZ-positive
cells in the first branchial arch of the double
mutants (Fig. 2, E and H). In contrast, other
embryonic sites of lacZ expression showed
similar staining in embryos of the different
genotypes. At E11.5, lacZ expression was observed in presumptive myoblasts within the first
and second branchial arches of normal embryos
(Fig. 2F). These lacZ-positive cells within the
first branchial arch surrounded the fifth cranial
nerve (Fig. 2J). There was a dramatic reduction
of lacZ staining in the myogenic cores of the
first branchial arches of double mutants at
E11.5 (Fig. 2, I and K), suggesting that the
capsulin-expressing myogenic lineage of the
first branchial arch was specifically affected in
these mutant embryos.
The striking lack of specific facial skeletal muscles in MyoR⫺/⫺capsulin⫺/⫺ double mutants was reminiscent of the phenotype associated with the combined absence
of Myf5 and MyoD, except that the latter
phenotype affects all skeletal muscles (2).
To define the temporal sequence of expression of these genes and to determine whether myogenic bHLH genes were expressed
in the affected muscle lineage of the double
mutant, we performed in situ hybridization
with adjacent sections of staged embryos
between E8.5 and 15.5. Capsulin transcripts were detected in mesodermal precursors of the first arch at E8.5 (Fig. 3A).
At this stage, Myf5, MyoD, and MyoR ex-
20 DECEMBER 2002 VOL 298 SCIENCE www.sciencemag.org
REPORTS
pression was undetectable. There has been
some disagreement about the timing of expression of Myf5 and MyoD in the branchial
arches, depending on the method of detection, but the earliest reported expression of
these genes in this region is E9.25 and
E9.5, respectively (22–24 ). By E9.5, Myf5
and capsulin were expressed in the same
cell population within the first branchial
arch, and by E10.5, Myf5, capsulin, and
MyoR were coexpressed in these cells of
wild-type embryos (Fig. 3A). In contrast,
Myf5 was not expressed in first branchial
arch precursors of MyoR⫺/⫺capsulin⫺/⫺
double mutants at E9.5 or E11.5 (Fig. 3B).
There was also no evidence for expression
of Myf5, MyoD, or myogenin at E15.5 in the
region of affected facial muscles (Fig. 3C),
whereas these genes were expressed in other developing head and trunk muscles.
To determine the fate of first arch muscle precursors that failed to activate expression of Myf5 and MyoD, we performed
TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling)
on histological sections of double-mutant
embryos at E10.5, when cells marked by
expression of capsulin-lacZ were disappearing. As shown in Fig. 4, TUNELpositive cells were observed among the
lacZ-positive muscle precursors of double
mutants, but not of MyoR⫺/⫺capsulin⫹/⫺
embryos. We conclude that these cells,
which fail to initiate the normal program
for muscle development in the double mutant, undergo apoptosis with resulting ablation of muscles of mastication. Similar
observations have been made in muscle
precursor cells in the limb buds of mice
lacking MyoD and myf5 (25).
The absence of specific head muscle cells,
as well as markers of the corresponding myogenic lineages, in MyoR⫺/⫺capsulin⫺/⫺ mutants resembles the effect of MyoD⫺/⫺Myf5⫺/⫺
double mutations on all skeletal muscles (2) and
is distinct from the phenotype of Myf5⫺/⫺
Pax3⫺/⫺ mutants, which exhibit a specific deficiency of trunk skeletal muscles (7 ). This
phenotype also differs from that of myogenin mutant mice, in which myoblasts express myogenic bHLH genes, but are unable to differentiate (3, 4 ). These findings
demonstrate that MyoR and capsulin redundantly regulate an initial step in the specification of a specific subset of facial skeletal
muscle lineages and that, in the absence of
these factors, myogenic bHLH genes are not
switched on, and cells from these lineages
undergo programmed cell death. There may
also be a modest effect on migration of precursors, as is seen in Lbx1 mutant mice (21).
MyoR and capsulin act as transcriptional
repressors in transfection assays (12, 20).
Whether they act to repress an inhibitor of
myogenesis or have a transcriptional-
activating function during development of
facial muscle remains to be determined.
The phenotype of MyoR⫺/⫺capsulin⫺/⫺
mutant mice reveals a previously unanticipated complexity in the development of
head skeletal muscles, and these findings
identify MyoR and capsulin as unique transcriptional regulators for the development
of specific head muscles.
References and Notes
1. M. Buckingham. Curr. Opin. Genet. Dev. 11, 440
(2001).
2. M. A. Rudnicki et al., Cell 75, 1351 (1993).
3. P. Hasty et al., Nature 364, 501 (1993).
4. Y. Nabeshima et al., Nature 364, 532 (1993).
5. A. Rawls, M. R. Valdez, W. Zhang, J. Richardson, W. H.
Klein, E. N. Olson. Development 125, 2349 (1998).
6. A. Rawls, E. N. Olson, Cell 89, 5 (1997).
7. S. Tajbakhsh, D. Rocancourt, G. Cossu, M. Buckingham, Cell 89, 127 (1997).
8. B. Christ, C. P. Ordahl, Anat. Embryol. (Berlin) 191,
381 (1995).
9. D. M. Noden, Am. J. Anat. 168, 257 (1983).
10. P. A. Trainor, S. S. Tan, P. P. Tam, Development 120,
2397 (1994).
11. A. Hacker, S. Guthrie, Development 125, 3461
(1998).
12. J. Lu, R. Webb, J. A. Richardson, E. N. Olson, Proc.
Natl. Acad. Sci. U.S.A. 96, 552 (1999).
13. L. Robb, L. Hartley, C. C. Wang, R. P. Harvey, C. G.
Begley, Mech. Dev. 76, 197 (1998).
14. J. Lu, J. A. Richardson, E. N. Olson, Mech. Dev. 73, 23
(1998).
15. H. Hidai, R. Bardales, R. Goodwin, T. Quertermous,
E. E. Quertermous, Mech. Dev. 73, 33 (1998).
16. S. E. Quaggin et al., Development 126, 5771 (1999).
17. L. Robb et al., Dev. Dyn. 213, 105 (1998).
18. J. Lu et al., Proc. Natl. Acad. Sci. U.S.A. 97, 9525
(2000).
19. S. E. Quaggin et al., Development 126, 5771 (1999).
20. J. Lu, E. N. Olson, unpublished results.
21. H. Brohmann, K. Jagla, C. Birchmeier, Development
127, 437 (2000).
22. M. O. Ott, E. Bober, G. Lyons, H. Arnold, M. Buckingham, Development 111, 1097 (1991).
23. S. Tajbakhsh, E. Bober, C. Babinet, S. Pournin, H.
Arnold, M. Buckingham, Dev. Dyn. 206, 291 (1996).
24. J. C. J. Chen, C. M. Love, D. J. Goldhammer. Dev. Dyn.
221, 274 (2001).
25. B. Kablar, K. Krastel, C. Ying, S. J. Tapscott, D. J.
Goldhammer, M. A. Rudnicki. Dev. Biol. 206, 21931
(1999).
26. We are grateful to C. Pomajzl and J. Stark for histologic preparations. We also thank A. Tizenor for
graphics and J. Page for editorial assistance. Supported by grants from the NIH, the Donald W. Reynolds
Foundation and the Muscular Dystrophy Association
to E.N.O.
Supporting Online Material
www.sciencemag.org/cgi/content/full/298/5602/2378/
DC1
Materials and Methods
Fig. S1
References
9 September 2002; accepted 28 October 2002
Genetic Structure
of Human Populations
Noah A. Rosenberg,1* Jonathan K. Pritchard,2 James L. Weber,3
Howard M. Cann,4 Kenneth K. Kidd,5 Lev A. Zhivotovsky,6
Marcus W. Feldman7
We studied human population structure using genotypes at 377 autosomal
microsatellite loci in 1056 individuals from 52 populations. Within-population
differences among individuals account for 93 to 95% of genetic variation;
differences among major groups constitute only 3 to 5%. Nevertheless, without
using prior information about the origins of individuals, we identified six main
genetic clusters, five of which correspond to major geographic regions, and
subclusters that often correspond to individual populations. General agreement
of genetic and predefined populations suggests that self-reported ancestry can
facilitate assessments of epidemiological risks but does not obviate the need
to use genetic information in genetic association studies.
Most studies of human variation begin by
sampling from predefined “populations.”
These populations are usually defined on the
basis of culture or geography and might not
reflect underlying genetic relationships (1).
Because knowledge about genetic structure
of modern human populations can aid in inference of human evolutionary history, we
used the HGDP-CEPH Human Genome Diversity Cell Line Panel (2, 3) to test the
correspondence of predefined groups with
those inferred from individual multilocus genotypes (supporting online text).
The average proportion of genetic differences between individuals from different human populations only slightly exceeds that
between unrelated individuals from a single
population (4–9). That is, the within-population component of genetic variation, estimated here as 93 to 95% (Table 1), accounts for
most of human genetic diversity. Perhaps as a
result of differences in sampling schemes
(10), our estimate is higher than previous
estimates from studies of comparable geographic coverage (4–6, 9), one of which also
used microsatellite markers (6). This overall
similarity of human populations is also evident in the geographically widespread nature
of most alleles (fig. S1). Of 4199 alleles
present more than once in the sample, 46.7%
appeared in all major regions represented:
Africa, Europe, the Middle East, Central/
www.sciencemag.org SCIENCE VOL 298 20 DECEMBER 2002
2381